What Matters 2 Adults: a study protocol to develop a new preference-based wellbeing measure with Aboriginal and Torres Strait Islander adults (WM2Adults)

Background Understandings of health and wellbeing are culturally bound. Many Aboriginal and Torres Strait Islander people perceive wellbeing and quality of life (QOL) differently from the Western biomedical models of health underpinning existing QOL instruments. Any instrument to measure the wellbeing of Aboriginal and Torres Strait Islander people should be culturally appropriate and safe, include relevant dimensions, and be informed by their own values and preferences. Existing QOL instruments do not meet these standards. This study will generate a new preference-based wellbeing measure, WM2Adults, for Aboriginal and Torres Strait Islander adults, underpinned by their values and preferences. Methods A mixed methods approach will be used; we will employ decolonising methodologies, privilege Aboriginal and Torres Strait Islander voices and perspectives, and adopt a strengths-based approach rather than a deficit lens. Yarning Circles will be conducted with Aboriginal and Torres Strait Islander people across Australia. A candidate item pool will be developed from these data, on which psychometric analysis and validity testing will be undertaken to develop a descriptive system. Following finalisation of the descriptive system, wellbeing states will be valued using a quantitative preference-based approach (best-worst scaling) with a diverse sample of Aboriginal and Torres Strait Islander adults (n = 1000). A multinomial (conditional) logit framework will be used to analyse responses and generate a scoring algorithm for the new preference-based WM2Adults measure. Discussion The new wellbeing measure will have wide applicability in assessing the effectiveness and cost-effectiveness of new programs and services for Aboriginal and Torres Strait Islander people. Results will be disseminated through journals, conferences and policy forums, and will be shared with Aboriginal and Torres Strait Islander communities, organisations and research participants.

Considering Fine-Grained and Coarse-Grained Information for Context-Aware Recommendations

In context-aware recommendation systems, most existing methods encode users’ preferences by mapping item and category information into the same space, which is just a stack of information. The item and category information contained in the interaction behaviours is not fully utilized. Moreover, since users’ preferences for a candidate item are influenced by the changes in temporal and historical behaviours, it is unreasonable to predict correlations between users and candidates by using users’ fixed features. A fine-grained and coarse-grained information based framework proposed in our paper which considers multi-granularity information of users’ historical behaviours. First, a parallel structure is provided to mine users’ preference information under different granularities. Then, self-attention and attention mechanisms are used to capture the dynamic preferences. Experiment results on two publicly available datasets show that our framework outperforms state-of-the-art methods across the calculated evaluation metrics.

A Knowledge-Aware Attentional Reasoning Network for Recommendation

Knowledge-graph-aware recommendation systems have increasingly attracted attention in both industry and academic recently. Many existing knowledge-aware recommendation methods have achieved better performance, which usually perform recommendation by reasoning on the paths between users and items in knowledge graphs. However, they ignore the users' personal clicked history sequences that can better reflect users' preferences within a period of time for recommendation. In this paper, we propose a knowledge-aware attentional reasoning network KARN that incorporates the users' clicked history sequences and path connectivity between users and items for recommendation. The proposed KARN not only develops an attention-based RNN to capture the user's history interests from the user's clicked history sequences, but also a hierarchical attentional neural network to reason on paths between users and items for inferring the potential user intents on items. Based on both user's history interest and potential intent, KARN can predict the clicking probability of the user with respective to a candidate item. We conduct experiment on Amazon review dataset, and the experimental results demonstrate the superiority and effectiveness of our proposed KARN model.

Novel Utility Procedure for Filtering High Associated Utility Items from Transactional Databases

In data mining, mining and analysis of data from different transactional data sources is an aggressive concept to explore optimal relations between different item sets. In recent years number of algorithms/methods was proposed to mine associated rule based item sets from transactional databases. Mining optimized high utility (like profit) association rule based item sets from transactional databases is still a challenging task in item set extraction in terms of execution time. We propose High Utility based Association Pattern Growth (HUAPG) approach to explore high association utility item sets from transactional data sets based on user item sets. User related item sets to mine associated items using utility data structure (UP-tree) with respect to identification of item sets in proposed approach. Proposed approach performance with compared to hybrid and existing methods worked on synthetic related data sets. Experimental results of proposed approach not only filter candidate item sets and also reduce the run time when database contain high amount of data transactions.

Hi-Fi Ark: Deep User Representation via High-Fidelity Archive Network

Deep learning techniques have been widely applied to modern recommendation systems, bringing in flexible and effective ways of user representation. Conventionally, user representations are generated purely in the offline stage. Without referencing to the specific candidate item for recommendation, it is difficult to fully capture user preference from the perspective of interest. More recent algorithms tend to generate user representation at runtime, where user's historical behaviors are attentively summarized w.r.t. the presented candidate item. In spite of the improved efficacy, it is too expensive for many real-world scenarios because of the repetitive access to user's entire history. In this work, a novel user representation framework, Hi-Fi Ark, is proposed. With Hi-Fi Ark, user history is summarized into highly compact and complementary vectors in the offline stage, known as archives. Meanwhile, user preference towards a specific candidate item can be precisely captured via the attentive aggregation of such archives. As a result, both deployment feasibility and superior recommendation efficacy are achieved by Hi-Fi Ark. The effectiveness of Hi-Fi Ark is empirically validated on three real-world datasets, where remarkable and consistent improvements are made over a variety of well-recognized baseline methods.

Convolutional Gaussian Embeddings for Personalized Recommendation with Uncertainty

Most of existing embedding based recommendation models use embeddings (vectors) to represent users and items which contain latent features of users and items. Each of such embeddings corresponds to a single fixed point in low-dimensional space, thus fails to precisely represent the users/items with uncertainty which are often observed in recommender systems. Addressing this problem, we propose a unified deep recommendation framework employing Gaussian embeddings, which are proven adaptive to uncertain preferences exhibited by some users, resulting in better user representations and recommendation performance. Furthermore, our framework adopts Monte-Carlo sampling and convolutional neural networks to compute the correlation between the objective user and the candidate item, based on which precise recommendations are achieved. Our extensive experiments on two benchmark datasets not only justify that our proposed Gaussian embeddings capture the uncertainty of users very well, but also demonstrate its superior performance over the state-of-the-art recommendation models.

Potential item set mining by using utility pattern growth model in big data

Information extracted by using data mining in earlier days. Now a day’s, the most talked about technology is Big Data. Utility Mining is the most crucial task in the real time application where the customers prefer to choose the item set which can yield more profit. Handling of large volume of transactional patterns becomes the complex issue in every application which is resolved in the existing work introducing the parallel utility mining process which will process the candidate item sets in the paralyzed manner by dividing the entire tasks into sub partition. Each sub partition would be processed in individual mapper and then be resulted with the final output value. The time complexity would be more when processing an unnecessary candidate item sets. This problem is resolved in the proposed methodology by introducing the novel approach called UP-Growth and UP-Growth+ which will prune the candidate item sets to reduce the dimension of the candidate item sets. The time complexity is further reduced by representing the candidate item sets in the tree layout. The test results prove that the proposed new approach provides better result than the existing work in terms of accuracy.

The Design and Analysis of the Information Management System Based on Data Mining

This paper studies on the data mining technology based on association rules, and analyzes on important algorithm in association rules - the advantages and disadvantages of Apriori algorithm and puts forward an improved Apriori-mapping algorithm based on address mapping. This algorithm adopts the way of horizontal deposit transaction, establishes candidate item identification list of corresponding candidate project transaction and length value of transaction list. And shorten the pruning operation time by address mapping, and compress the frequent item sets number of operation connected operation with large amplitude.The system efficiency is improved, and the performance of the algorithm has been improved by experiment.